SPI，TFT触摸屏与PIC18结合在一起

汤姆，为什么这么急？你问过同样的问题两次吗？



      以下为原文

    Tom, What makes this urgent ?
Did you ask the same question twice ?

嗨，吉姆，好像我贴了两次。我是新来的论坛。如何删除其他帖子？



      以下为原文

    Hi Jim,
It seems I posted twice. I'm new to the forum. How do I delete the other post?

最好的方法是编辑另一个。删除内容/写入重复。



      以下为原文

    The best way is to edit the other one. Delete content/ write duplicate.

我目前正在开发一个在杂货店使用的系统。对于用户界面，我使用的是连接到PIC的7个“River界面”。我一直在用FTDI提供的库编写应用程序。基本上建立在他们的应用程序之上。代码导致River屏幕在一段时间后冻结。我找到了一种方法，使程序员通过阅读程序员指南的第5.6部分（第157页）来恢复屏幕。http://btcx.com /Wp CON…sjava程序员指南。PDFSO我在我们的应用程序中应用了这个。通过检测最大迭代次数，检测代码是否被困在while循环中。在这种情况下，我执行恢复过程。然而，执行恢复过程并不理想。这是因为文本在发生的几秒钟内变得混乱，这对用户来说是显而易见的。更重要的是，屏幕最终会关闭或变空，直到屏幕关闭电源。这可能是因为恢复过程太频繁了。我认为找到问题的根源比修补要好得多。这样，恢复过程根本不需要执行，屏幕也不会崩溃。我必须在4月6日（用PCB）演示产品。正如现在，屏幕可能工作长达20分钟，但它可能随时崩溃。我想解决这个问题的演示日期，使演示顺利进行。



      以下为原文

    I'm currently developing a system to be used in grocery stores. For the user interface, I'm using the 7" Riverdi screen connected to the PIC.
 
I've been writing the application with the library provided by FTDI. Essentially building on top of their application.
 
The code causes the Riverdi screen to freeze after some time. I found a way to make the screen recover by reading section 5.6 (p. 157) of the programmer's guide.
http://brtchip.com/wp-con...s_Programmer_Guide.pdf
 
So I applied this in our application. I detect if the code is stuck at the while loop by having a max number of iterations. In that case I execute the recovery procedure. However, executing the recovery procedure is not ideal. This is because the text becomes garbled for a few seconds when it happens, which is noticeable to the user. What's more, the screen eventually turns off or goes blank and it stays that way until the screen is powered off and on. This may be because the recovery procedure is being done too often.
 
I believe that finding the root of the problem would be a better approach than patching. This way the recovery procedure would not need to be executed at all and the screen would not crash.
 
I have to demonstrate the product on April 6 (with a PCB). As it is right now, the screen may work for up to 20 minutes but it may crash any time. I would like to resolve this issue by the date of the demo so that the presentation goes well.

勘误表说有一个SPI主模式问题。所以检查错误是否影响显示应该是下一个步骤。7。模块：MSSP（主同步串行端口）7.1 SPI主模式，当MSP在SPI主模式中使用，并且CKE位是清晰的（CKE＝0）时，缓冲器满（BF）位和MSSP中断标志（SSPIF）位早就设置为半SCK周期。如果用户软件立即对所设置的任一位作出反应，则会发生写入冲突，如WCOL位所指示的。为了避免写入冲突，应使用以下方法之一：方法1：在检测到完成的传输（BF位或SSPIF位设置）之后，在写入SSPBUF寄存器之前，添加一个SCK周期的软件延迟。在写入SSPBUF后，验证WCOL位是否清晰。如果设置了WCOL位，则清除软件中的位并重写SSPBUF寄存器。方法2：作为MSSP初始化过程的一部分，设置CKE位（CKE＝1）。受影响的硅修订A2 A4×A2 A4×XHM，似乎使用方法2，CKE为1…



      以下为原文

    The Errata says there is an SPI Master mode issue.
 
So checking if that error effects the display should be the next step.
 
7. Module: MSSP (Master Synchronous Serial Port) 7.1 SPI Master mode When the MSSP is used in SPI Master mode and the CKE bit is clear (CKE = 0), the Buffer Full (BF) bit and the MSSP Interrupt Flag (SSPIF) bit become set half an SCK cycle early. If the user software immediately reacts to either of the bits being set, a write collision may occur as indicated by the WCOL bit being set. Work around To avoid a write collision one of the following methods should be used:
 
Method 1: Add a software delay of one SCK period after detecting the completed transfer (the BF bit or SSPIF bit becomes set) and prior to writing to the SSPBUF register. Verify the WCOL bit is clear after writing to SSPBUF. If the WCOL bit is set, clear the bit in software and rewrite the SSPBUF register.
Method 2: As part of the MSSP initialization procedure, set the CKE bit (CKE = 1).
 
Affected Silicon Revisions A2 A4 X A2 A4 X X
 
hm, seems method 2 is used, CKE is 1...

应用程序在使用River 5屏幕时不会冻结。所以我们选择了这一点。可能是购买的7个屏幕中的一个缺陷？



      以下为原文

    The application doesn't freeze when using a Riverdi 5" screen. So we opted for that. Maybe a defect in the 7" screen that was purchased?